ISSN: 2456-7132
Volume 8, Issue 1, pp. 143-149, 2020
DOI: https://doi.org/10.21467/ias.8.1.143-149
Copyright © 2019. The Author(s). Published by AIJR Publisher.
This is an open access article under Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license,
which permits any non-commercial use, distribution, adaptation, and reproduction in any medium, as long as the original work is
properly cited.
S H O R T R E V I E W
Steganographic Techniques Classification According to Image Format
Khaldi Amine
Computer Science Department, Faculty of Sciences and Technology, Artificial Intelligence and Information
Technology Laboratory (LINATI), University of Kasdi Merbah, 30000, Ouargla
* Corresponding author email: amine.vision@live.fr
Received: 05 June 2019 / Revised: 24 October 2019 / Accepted: 28 October 2019 / Published: 04 November 2019
A B S T R A CT
In this work, we present a classification of steganographic methods applicable to digital images. We
also propose a classification of steganographic methods according to the type of image used. We
noticed there are no methods that can be applied to all image formats. Each type of image has its
characteristics and each steganographic method operates on a precise colorimetric representation. This
classification provides an overview of the techniques used for the steganography of digital images
Keywords: Steganography, data hiding, digital image, Least Significant Bit, DCT, LUT.
1 Introduction
The problem of the secret data exchange has
always existed. Cryptography provides an
effective means of protecting secret data by
making it unintelligible to unauthorized persons;
however, the simple act of communicating with
encrypted messages attracts attention. This can
be problematic when it concerns a
communication channel monitored by a third
party, which can, at the slightest suspicion,
destroy the communication between the two
parties [1]. In these cases, a communication
containing a secret message between two persons
should appear normal to the person controlling
the channel. For this scenario, steganography
represents the alternative to cryptography.
Steganography, or the science of secret
communication, is a method of hiding a secret
message within an innocuous host medium, so
that the resulting medium appears to be
unaffected (undetectable concealment) by
inserting the secret message. The goal is to go
unnoticed a message in another message. It is
distinguished from cryptography, which seeks to
make a message unintelligible to other than
whomever it may concern [2]. The aim is to make
it difficult or impossible to distinguish between
an original medium and a medium modified by
the insertion of a secret message. Nowadays, with
the development of the Internet and the
explosion of digital mediums (sounds, images,
and videos) shared on the different
communication networks, steganography
becomes a popular practice and accessible to
anyone wishing to communicate discreetly with
other people. Steganography is based on the idea
of security by obscurity: if no one knows that
there is has a hidden file, no one will look at it or
retrieve it. And with everything that goes on the
Internet, and the number of attached files that
people are exchanging, nobody has enough
computer resources to scan all these transfers of
images, sounds and other files [3]. The scientific
community was particularly interested in this
discipline. The researchers showed that the
steganography applied to the current digital
media represents a real challenge involving many
disciplines: mathematics, statistics, signal
processing, information theory, and game theory.
Among the mediums which are very suitable for
the concealment of information, we distinguish
digital images. Since this type of medium is very
commonly exchanged on the Internet, a great
majority of the research work is devoted to it. In
this paper, we are also interested in digital images
as a medium cover. We will in this paper make a
classification of steganographic methods and
144
ISSN: 2456-7132
Available online at Journals.aijr.in
Steganographic Techniques Classification According to Image Format
those according to the digital image formats on
which they operate. The remainder of the paper
is organized as follows: In section 2 we present
the steganography and these main characteristics
and property. Then, in section 3 we will detail the
different steganographic techniques and
methods. In the fourth and last section we
propose a classification of steganographic
methods.
2 Steganography
Steganography (steganos Greek, cover and
graphein, writing) is the art of hiding a secret
message within another carrier message (text,
image, sound, video …) of an innocuous
character, so that the existence of the secret is
hidden from it. With cryptography, security is
based on the fact that the encrypted message is
incomprehensible to unauthorized persons, with
steganography, security is based on the fact that
the presence of a secret message will probably not
be suspected and detected. Inserting a message
into the chosen file involves changing parts of his
code. The whole art of steganography is to make
sure that these changes are invisible or inaudible
[4]. The smaller the message and the larger file,
the more this alteration is likely to go unnoticed.
2.1 Steganography Characteristics
The goals of information concealment can
change in a subtle way [5]. Classically,
applications are sorted according to three criteria:
• Robustness ensures that secret
information cannot be destroyed
without severely degrading the image.
It quantifies the resistance of the
concealed message to the various
attacks (transformations) made to the
stego-medium.
• The invisibility aims to ensure that the
stego-image is not disturbed by the
inserted secret information.
• The insertion capacity of a
steganography system is defined by
the size in bits of the secret message
which can be integrated into a
medium of given size. The relative
insertion capacity is the ratio between
the size of the secret message to be
concealed and the size of the medium
used. Capacity therefore defines the
amount of information that can be
integrated into the medium without
visible deterioration.
These three characteristics are closely related
and inverse (Figure 1). For example, capacity
improvement generally has a negative
influence on invisibility
Figure 1: Steganography characteristics
2.2 Steganography protocols
Pure steganography: no prior agreement,
other than choice of algorithm is necessary, A
and B use the channel to exchange
information [6] (Figure 2).
Figure 2: Pure steganography process
Steganography with secret key: A and B
agree beforehand of a key used to insert and
then extract the message of the stego-medium
[7] (Figure 3).
Figure 3: Secret key steganography process
Public key steganography: like in
cryptography, A uses the public key of B when
it wants to send a message to it. B extract with
its private key [8] (Figure 4).
Figure 4: Public key steganography process
145
ISSN: 2456-7132
Available online at Journals.aijr.in
Khaldi Amine, Int. Ann. Sci.; Vol. 8, Issue 1, pp: 143-149, 2020
2.3 Areas of steganography
Steganography is divided into two domains,
spatial and frequency. In the spatial domain,
the secret message is inserted into the pixels
of the carrier image, whereas in the frequency
domain the pixels are transformed into
coefficients, and the secret message is inserted
in these coefficients [9].
Spatial domain: Spatial steganography
involves changing bits of pixels in the image
to insert the bits of the secret message. The
LSB technique is one of the simplest and most
common techniques. It consists in hiding a
secret message in the least significant bits of
the pixels of the image, so that the distortions
brought by the insertion process remain nonperceptible. The reason is that for the human
eye, variations in the value of the LSB are
almost imperceptible. The insertion of secret
message bits may be done sequentially or
pseudo randomly.
Frequency domain: The message is inserted
into the transformed coefficients of the image,
which has the effect of bringing more
robustness against the attacks. Frequency
steganography is an essential technique for
concealing secret information: nowadays most
steganography systems operate in the
frequency domain. The frequency
steganography will thus make it possible to
hide the information in areas of the image less
sensitive to compression, cropping and
various image processing.
2.4 Classification of Steganographic
Methods (Image Format)
Before being able to establish our
classification, it is essential to present the
types of images chosen for our classification.
We have chosen the most commonly used
formats that are currently used and
exchanged. Different representations of
images must be distinguished. In a file, for
storing and exchanging data, the image is
usually compressed and stored in a graphic
format. The main matrix formats are Windows
bitmap (BMP), Graphics Interchange Format
(GIF), Portable Network Graphics (PNG) and
Joint Photographic Experts Group (JPEG).
Each format has its own characteristics. To
choose the one that corresponds to what we
want to do with our images, it is essential to
know the depth of the colors. Expressed in
bits, it corresponds to the number of color
values that each pixel of the image can take.
• BMP (or Bitmap) format is a universal,
uncompressed format developed by
Microsoft and IBM. It allows a faithful
reproduction of the colors of the original
image; the counterpart is the high weight
of the generated file.
• Graphics Interchange Format (GIF):
Also widely used in the web, this
proprietary format uses an indexed color
scheme. It is thus possible to use only
specific color values, which makes it
possible to optimize as much as possible
the weight of the visual. In contrast, the
appearance of visuals displaying many
colors is strongly degraded. This format
also allows you to create frame-by-frame
animations.
• The Joint Photographic Experts Group
(JPEG) format: widely used in the web,
this format was developed by a panel of
experts that publishes compression
standards for still images. This
compressed format greatly alters the
quality of the original images but allows
a relatively accurate color reproduction
and a relatively light weight.
• Portable Network Graphics (PNG)
format: This format is standardized by
the World Wide Web Consortium
(W3C) and is designed to bypass the
existing proprietary GIF format. The
PNG thus has exactly the same
characteristics as the GIF, and also
allows the recording from 1 to 48 bits,
and the management of transparency
(alpha channels). On the other hand, it is
not possible to make animations.
146
ISSN: 2456-7132
Available online at Journals.aijr.in
Steganographic Techniques Classification According to Image Format
2.5 Classification of steganographic
techniques
As we can see in the figure 5, the steganographic
methods can be studied in five groups: Fusion
methods, Statistical modification technics,
permutation, substitution and additive marking
technics.
Figure 5: Steganographic techniques classification
Fusion: This technique, which can be considered
as naive steganography [9], consists in adding the
data to be hidden to the file. To do this, this
method uses unused or unread slots by most
image decoders. As we can see in “Figure 6”
There are two operations: Adding data at the end
of the file and adding to the file headers.
The addition at the end of the file is
made possible by the fact that most
image decoders do not read the image
file as a whole. For most available image
formats, a certain bit string is set to mark
the end of the image.
The addition at the end of the image
simply appends the hidden data after this
string. No size limitations are imposed;
however, a 20 Mbytes image file may not
go unnoticed. As for adding to the
header, some formats such as the bitmap
define a field to specify the offset from
which the image will start. By specifying
a slightly longer offset it is possible to
hide the data to be concealed between
the two offsets.
Statistical methods: Statistical methods modify
several support statistics (letter frequencies, pixel
distribution) to hide the message [10] and retrieve
it by testing these assumptions. In this technic,
the generated mark is directly inserted into the
original image. The process of detection and
extraction, in this type of steganography (for
blind techniques) is carried out using statistical
methods. For example, a correlation
measurement can be performed.
Additive Schemes: During insertion, the signal
representing the mark is added to certain
components of the medium. In order to do this,
it is neither necessary to adapt the mark to the
medium so that the signal it represents is not too
low (risks of non-detectability) nor problems of
robustness nor too strong (erasure of the initial
signal and therefore too great degradation of it).
Spectrum spreading [11] is a technique used in
radio telecommunications, especially by the
military, to disperse a signal over a wide
frequency band to make it discreet and resistant
to interference.
Figure 6: Classification of steganographic techniques according to the image format
147
ISSN: 2456-7132
Available online at Journals.aijr.in
Khaldi Amine, Int. Ann. Sci.; Vol. 8, Issue 1, pp: 143-149, 2020
Figure 7: Substitution technics classification
Substitution technics: In substitution methods,
the information to be hidden is not added but
substituted in the components of the image
(pixel, transform coefficient) selected using a
secret key. As we can see in “Figure 7”
substitution technics may be subdivided into two
sub-categories, direct and indirect. In the direct
substitution the cover values are modified
without any transformation process. In the
indirect substitution, a transformation of the
cover is made before the concealment process.
• Bit Plane Complexity Segmentation
(BPCS): The Human visual system has
such a special property that a toocomplicated visual pattern cannot be
perceived as “shape-informative” [12] For
example, on a very flat beach shore every
single square-foot area looks the same – it
is just a sandy area, no shape is observed.
However, if you look carefully, two samelooking areas are entirely different in their
sand particle shapes. BPCSSteganography makes use of this
property. It replaces complex areas on the
bit-planes of the vessel image with other
complex data patterns (i.e., pieces of
secret files). This replacing operation is
called “embedding.” No one can see any
difference between the two vessel images
of before and after the embedding
operation.
• Modification of DCT coefficients: This
approach consists in extracting a certain
number of squares of 8 × 8 pixels from
the image, calculating the DCT transform
of these blocks and marking a bit on the
averages frequencies, the modification of
the low frequencies of the image would
not change it too much. The low
frequencies corresponding to the largest
homogeneous areas in the image, for
example uniform black in the dark areas,
and the high frequencies being removed
by JPEG compression, corresponding to
the smallest homogeneous areas of an
image [13].
• Least Significant Bit (LSB): LSB
gathers everything related to data
concealment by modifying the low-order
bit of an element [14]. Modification of the
value of a pixel or the modification of the
value of a DCT coefficient in the case of
the JPEG standard. All are based on the
insensitivity of the human visual system to
a small change of colors. There are two
LSB methods:
LSB replacement: This technique
consists in substituting the least
significant bits of the pixels for the
message bits to be inserted. To insert a
message, the last least significant bit of
each pixel is replaced by a bit of the
148
ISSN: 2456-7132
Available online at Journals.aijr.in
Steganographic Techniques Classification According to Image Format
message to be concealed. The path
direction of the pixels is usually chosen by
a pseudo-random path. To do this, the
transmitter and the receiver must first
exchange a key k, used as the seed of a
pseudo-random number generator.
LSB matching steganography: Least
Significant Bit (LSB) matching
steganography, also named ±1
embedding, is a slightly more
sophisticated version of LSB embedding.
The LSB correspondence steganography
method does not alter the first order
statistical distribution of the host support.
So All first-order statistical attacks are
ineffective.
As we can see in the figure 7, a steganographic
technique cannot always be applied to all image
formats. It is not possible, for example, to modify
the DCT coefficients of a BMP image since this
format is not compressed, it is also not possible
to modify the LUT in a BMP image because the
colorimetric representation in a BMP image does
not use Lookup table. For JPEG, PNG and GIF
images, direct substitution cannot be performed
as both types require processing to access pixel
values (DCT for JPEG and LUT for GIFs).
3 Conclusion
The need for secret or discrete communication is
not a new quest: since antiquity, human beings
have always sought to protect and disguise their
data with different methods. With the advent of
the Internet, adapted digital methods were then
put in place. In this paper, we are interested in
steganography which is a secret communication
process. In our work we have presented the most
used steganographic techniques, we have also
tried to classify these techniques according to the
types of images on which they are applied. In our
classification we also distinguish steganography
techniques depending on the way the message is
inserted into a host document, either additively
known by additive techniques, or alternatively
known by substitution techniques. Insertion in
the spatial domain has logically been the first to
be considered. But, if the methods used are quite
simple, they often show their limits fairly quickly.
The spatial domain has the advantage of being
inexpensive in computing time, since it is not
necessary to perform transformations. However,
this domain does not easily handle invisibility. In
addition, it is natural to think that steganography
is incompatible with lossy compressions. In
reality, it is much more advantageous to use the
JPEG format, which is much more used and
allows more discretion, even if it is done at the
expense of the insertion capacity. We note in our
classification that there are no methods
applicable to all image formats, each format has
its characteristics, this classification could be
useful because it gives an overall view of the
existing methods and applicable to digital images
and it allows delimiting the fields of application
of each technique. The extension of this work
could lead to a wider classification by integrating
all possible digital media (audio and video).
4 Competing Interests
The author declared that no conflict of interest
exists in the publication of this work.
How to Cite this Article:
A. Khaldi, “Steganographic Techniques Classification
According to Image Format”, Int. Ann. Sci., vol. 8, no. 1,
pp. 143-149, Nov. 2019. doi: 10.21467/ias.8.1.143-149
References

International Conference on Computing for Sustainable
Global Development, India, 2014.

K.Jung, “Image steganography in spatial domain: A
survey”, Signal Processing: Image Communication,
Volume 65, Pages 46-66, July 2018.

identification”, Defense Technology, Volume 15, Issue
2, April 2019, Pages 162-169.

Forensics and Security, Volume 10, Issue 2, April 2015,
Pages 243 – 255.
149
ISSN: 2456-7132
Available online at Journals.aijr.in
Khaldi Amine, Int. Ann. Sci.; Vol. 8, Issue 1, pp: 143-149, 2020
[7] S.Dagar, “Highly randomized image steganography
using secret keys”, International Conference on Recent
Advances and Innovations in Engineering, India, 2014.
[8] Z.Li, Y.He, “Steganography with pixel-value
differencing and modulus function based on PSO”,
Journal of Information Security and Applications,
Volume 43, December 2018, Pages 47-52.
[9] A. El_Rahman, “A comparative analysis of image
steganography based on DCT algorithm and
steganography tool to hide nuclear reactors confidential
information”, Computers & Electrical Engineering,
Volume 70, August 2018, Pages 380-399.
[10] A.Shaik, V. Thanikaiselvan R.Amitharajan, “Data
Security Through Data Hiding in Images: A Review”,
Journal of Artificial Intelligence, Volume 10, 2017,
Pages 1-21.
[11] M.Nutzinger, C.Fabian, M.Marschalek, “Secure Hybrid
Spread Spectrum System for Steganography in Auditive
Media”, International Conference on Intelligent
Information Hiding and Multimedia Signal Processing,
Germany, 2010.
[12] S.Bhattacharyya, A.Khan, A.Nandi, A.Dasmalakar,
S.Roy, G.Sanyal, “Pixel mapping method (PMM) based
bit plane complexity segmentation (BPCS)
steganography”, World Congress on Information and
Communication Technologies, India, 2011.
[13] K.B. Raja, C.R. Chowdary, K.R. Venugopal, L.M.
Patnaik, “A Secure Image Steganography using LSB,
DCT and Compression Techniques on Raw Images”,
Third International Conference on Intelligent Sensing
and Information Processing, India, 2005.
[14] G.Swain, “Two new steganography techniques based on
quotient value differencing with addition-subtraction
logic and PVD with modulus function”, Optik, Volume
180, February 2019, Pages 807-823.
Publish your books with AIJR publisher-
✓ Publish with ISBN and DOI.
✓ Publish Thesis/Dissertation as Monograph.
✓ Publish Book Monograph.
✓ Publish Edited Volume/ Book.
✓ Publish Conference Proceedings
✓ Retain full copyright of your books.
Submit your manuscript at books.aijr.org
Publish your research article in AIJR journals-
✓ Online Submission and Tracking
✓ Peer-Reviewed
✓ Rapid decision
✓ Immediate Publication after acceptance
✓ Articles freely available online
✓ Retain full copyright of your article.
Submit your article at journals.aijr.in